As enterprises increasingly adopt autonomous AI agents, iProov has introduced a new human approval service designed to ensure that high-risk actions remain subject to trusted human oversight.
The service implements the Human Approval and Presence Specification (HAPS), released in September 2026, providing cryptographically verifiable proof that the genuine, authorized individual approved a specific action before an AI agent can proceed. It integrates with agentgateway, the open-source AI agent gateway developed by solo.io, and Cerbos, an authorization management platform and the service’s first policy-engine partner.
The new HAPS engine combines proof of human identity and presence through iProov’s Dynamic and Express liveness technologies, which are used by national governments and major organizations millions of times daily worldwide. It connects this verification to approval of a specific action, generating an auditable, non-repudiable certificate for logging and future reference.
Bringing Verifiable Human Oversight to Agentic AI
The launch comes as industry leaders increasingly emphasize the importance of human oversight in autonomous AI systems. In an October 2026 Microsoft Security blog, Microsoft highlighted the need for policies that can “require cryptographically provable human approval for a high-impact step.”
The integration of iProov’s Hosted HAPS Approval Engine with Cerbos and agentgateway addresses this emerging security requirement. It enables organizations to verify that the appropriate human has approved a specific action before an AI agent is authorized to execute it.
“Cerbos controls and secures what decisions AI Agents can make in an organization, and under what conditions,” said Alex Olivier, Chief Product Officer and Co-Founder of Cerbos. “Our partnership with iProov introduces a new level of control, enabling organizations to require approval from the right human before a high-consequence action can proceed. By adding this decision capability, we are giving organizations the confidence to delegate an even greater range of tasks to their AI agents than ever before while still retaining human oversight and accountability.”
Andrew Bud, founder and CEO of iProov, emphasized that organizations must balance the growing capabilities of AI agents with clear human accountability.
“For organizations to realize the full potential of agentic AI, they need to task agents with increasingly consequential responsibilities. But some actions must remain subject to human control, approval and accountability,” said Bud. “Our new HAPS service provides cryptographically verifiable proof that the real, right human being, confirmed by iProov Genuine Human Presence, has approved a specific action. Cerbos determines when such human approval is required and calls the HAPS service. Together, we enable organizations to embrace AI agent automation without compromising human control. By requiring a genuine human face as part of the approval process, we secure the process against agent impersonation. Because no matter how smart they become, agents will never have faces.”
How the HAPS Approval Process Works
When a Cerbos policy determines that an action requires human approval, the AI agent’s request is temporarily held while iProov’s hosted HAPS service is activated.
The service presents the designated approver with the exact action and its scope. If the individual approves, they complete an iProov genuine human presence check. The HAPS engine then issues a signed, single-use credential tied specifically to the approved action, allowing the agent to proceed.
The credential is valid only for the exact details presented during approval. If any information changes, such as the amount or destination of a payment, the credential no longer matches and the action is blocked. This creates verifiable evidence that the appropriate person approved the specific action, rather than relying on an AI agent to authorize its own request.
An Open Specification for Wider Adoption
The HAPS specification remains open, allowing any gateway, policy engine, or relying party to implement compliant human approval workflows. Implementations can be supported by iProov or other forms of proof.
The open-source HAPS specification, a partial Rust reference implementation, and test vectors are available on GitHub under the Apache-2.0 license.
Dominic Forrest, Chief Technology Officer at iProov, is set to present the HAPS service and its integration with agentgateway and Cerbos at AgntCon North America, scheduled for October 22–23 in San Jose, California.
With its new service, iProov aims to help enterprises expand AI agent automation while maintaining verifiable human control over consequential decisions.
